What Is Local AI?
AI that runs on your own device, not on a company’s servers.
The simple definition
With cloud AI, your message travels to a data center, a large model answers, and the answer travels back. ChatGPT, Gemini, and Claude normally work this way. With local AI, the model is on your device and nothing has to leave it.
Cloud AI = your question goes to a server. Local AI = your question never leaves your device.
Different names, same idea
- Local AI, on-device AI: the general terms.
- Local LLM: the same thing, as developers say it.
- Offline AI: named for the result. It works with no connection.
- Edge AI: the industry's word for it.
- Private AI: a claim, not a design. A cloud app can say it too.
Why it works now
Models have been compressed to a fraction of their size while keeping most of their ability, and phones now ship with chips built to run them.
What it's good for
- Privacy: no server sees your conversation.
- Working offline: no Wi-Fi, no signal, no outages. See why offline AI matters.
- Cost: no server bills, so often no subscription.
- Speed on simple tasks: no network round-trip.
What it can't do (yet)
Local models are smaller than the largest cloud models, so they are not the best choice for very complex reasoning, huge documents, or anything that needs live information. See offline AI vs. cloud AI.
In practice: Cloak
Cloak: Offline AI is an AI chat app for iPhone and iPad built by Polyaura. It downloads its model once, then runs it on your device: no server sees your messages, and no account or connection is needed. It is on the App Store.
Cloak: Offline AI
Chat, Voice Memos, and Recap. $19.99 once.
